Skip to main content

Table 1 Simulation scenarios

From: Robust estimation of the expected survival probabilities from high-dimensional Cox models with biomarker-by-treatment interactions in randomized clinical trials

Scenarios

Effect size

Censoring rate

α

β j

γ j

T−

T+

(1) Complete null

0

0

0

0.72

0.72

(2) Treatment effect only

− 0.8

0

0

0.62

0.80

(3) 20 prognostic markers

0

~ U(−0.05, − 0.20)

0

0.70

0.70

(4) 15 treatment-effect modifiers

0

0

~ U(−0.10, − 0.40)

0.71

0.71

(5) Treatment effect + (4)

− 0.8

0

~ U(−0.10, − 0.40)

0.61

0.78

(6) 20 prognostic markers + (5)

− 0.8

~ U(−0.05, − 0.20)

~ U(−0.10, − 0.40)

0.60

0.76

  1. T+: experimental arm, T−: control arm